Pytorch 기본 문법 이해

data1 = torch.FloatTensor([[0, 1, 2, 3, 4], 
[5, 6, 7, 8, 9], 
[10, 11, 12, 13, 14]])
print(data1, data1.shape)
print(data1[1,2])
print(data1[:, 1])
print(data1[:,-1])
print(data1[:, :]) #전체

data1 = torch.FloatTensor([[0, 1, 2, 3, 4], 
[5, 6, 7, 8, 9], 
[10, 11, 12, 13, 14]])
data2 = data1 > 7
print(data2) #T/F 출력
print(data1[data2]) #T인 값만 추출

print(data1[(data1 < 3)| (data1 > 11)])

data1 = torch.randn(4,3)
print(data1)
print(data1[[1,2]])
print(data1[[2,1]])

print(data1[[1,2]][:, [0,2]])

data1 = torch.FloatTensor([[1, 2, 3],[4, 5, 6]])
print(data1)
data2 = data1[:2, :2]
data2[1, 1] = 4
print(data2)
print(data1)

data1 = torch.ones(5, requires_grad=True)
print(data1)

data2 = data1.clone()
data2[1] = 2
print(data1, data2.requires_grad)

data3 = data1.detach()
data3[1] = 3
print(data1, data3.requires_grad)

data1 = torch.ones(5, requires_grad=True)
print(data1)

data2 = data1.clone().detach()
data2[1] = 2
print(data1, data2.requires_grad)

data3 = data1.detach().clone()
data3[1] = 3
print(data1, data3.requires_grad)

data1 = torch.FloatTensor([7, 2, 0, 4, 1])
index = torch.where(data1 < 3)
print(index)
print(data1[index])

data2 = torch.where(data1 < 3, -1, 1)
print(data2)

data1 = torch.FloatTensor([ [1, 2, 3], [4, 5, 6] ])
print (data1.min())
print (data1.max())
print (data1.sum())
print (data1.mean())
print (data1.var())
print (data1.std())
print (data1.argmin()) #인덱스
print (data1.argmax()) #인덱스

data1 = torch.linspace(1, 5, 4)
print(data1)
torch.save(data1, 'mydata1.pt')

data2 = torch.load('mydata1.pt')
print(data2)

data1 = torch.linspace(1, 5, 4)
data2 = torch.linspace(1, 10, 4)
data3 = torch.linspace(1, 100, 4)
print (data1)
print (data2)
print (data3)

datas1 = {'data1': data1, 'data2': data2, 'data3': data3}
torch.save(datas1, 'mydata2.pt')

datas2 = torch.load('mydata2.pt')
print (datas2['data1'])
print (datas2['data2'])
print (datas2['data3'])